Statistical models can represent complex relationships among pieces of data. These models can be generated from large data sets that have hundreds or thousands of interrelated pieces of data and can be very challenging to formulate. It can take a significant amount of time, memory, processing power, and electrical power to generate these statistical models. For example, a computing device may perform hundreds of thousands of calculations over hundreds of iterations of steps to generate these models, during which time the computing device may repeatedly access and store information in thousands of memory locations. This level of processing can take hours, or days, to perform; can require a significant amount of electrical power; can reduce the available resources (e.g., processing power and memory) on the computing device for performing other tasks; can slow down other processes executing on the computing device; and can require complicated and expensive hardware.